Gather More Than You Use

Research goes wrong in two opposite directions from the same mistake, the belief that the amount gathered and the amount used should match. One writer treats research as a debt owed to the reader: three weeks went into the guild system, so the guild system goes in, and the reader gets proof of effort instead of a work. Another gathers too little, then has nothing to choose among and fills the pages with whatever there is.

When John McPhee’s editor at The New Yorker returned his forty-thousand-word manuscript about oranges, McPhee recalls, the envelope was large but slimmer than a postcard. Robert Bingham had condensed the piece. McPhee took a train into the city and told William Shawn, the magazine’s editor, that eighty-five percent of what he had written was gone. After five days of talk with Bingham, enough was restored to run the piece in two issues, but not all of what had been cut. McPhee had meant to write something under ten thousand words, until a visit to a citrus research station in Florida turned up several dozen people with doctorates in oranges and a library of a hundred thousand titles. The library, he concludes, had beguiled him so thoroughly that he “lost the advantage of what is left out.”

Gather generously anyway, because the shape of a topic appears only when more of it is on the table than the piece can hold, and then select for the reader, who is owed a work and not a record of your acquisitions. You needn’t destroy what you cut, and you may not drop what complicates your claim; in nonfiction, selection that removes the disconfirming case distorts the account (Practice 73). The reader’s promise (Practice 16) supplies the standard, but a standard can only be applied to a pile large enough to refuse.

Robert Caro met the same surplus by enlarging the container. According to The Guardian, his editor Robert Gottlieb cut 350,000 words from The Power Broker, which is why the Johnson biography was planned in several volumes, so that nothing major would have to go.

Practice. Find one fact, example, scene, or digression in your current draft that you find interesting, the kind you would tell a friend about, and ask what the reader would lose if it vanished. Move it to a separate file, read the passage without it, and then decide in writing, within fifteen minutes of starting: gone, restored, or moved elsewhere. Notice whether the passage became more or less sure of itself. If nothing was lost, you have also learned something about the edge of the Work.